JP3873299B2 - Kana-kanji conversion device and kana-kanji conversion method - Google Patents

Description

[0001]
[Industrial application fields]
The present invention relates to a kana-kanji conversion device and a kana-kanji conversion method, and more particularly to a kana-kanji conversion device and a kana-kanji conversion method that use dependency information between words for phrase segmentation or word kanji candidate selection.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, various kana-kanji conversion devices that convert a kana character string input from a keyboard or the like into a desired kana-kanji mixed sentence have been proposed as a Japanese sentence input device or a Japanese sentence editing apparatus. As such a kana-kanji conversion device, it is not necessary for the user to designate the position of separating words and phrases one by one, and it is desired that the converted character string has the notation desired by the user. In Japanese, there are many homonyms and homonyms, so in order to obtain the correct kana-kanji mixed sentence without error, the meaning of the sentence must be analyzed in the end. In order to analyze, it is necessary to have a knowledge base of tens of thousands of words that are at least organically related.
[0003]
Therefore, the conventional kana-kanji conversion device attempts to obtain the result desired by the user without analyzing the meaning by devising the phrase segmentation process and the learning process for selecting the homonym. The segment segmentation process includes a two-segment longest match method in which a longest phrase is obtained as a first candidate among phrases that can be composed of two phrases as a basic unit, a word that can be a candidate for a word constituting the phrase, and There is a minimum cost method in which a cost is given to a combination of words, and a phrase whose score satisfies a predetermined condition is a first candidate. Also, in the learning process, learning of homophones with the highest priority given to the word selected by the user immediately prior to the homophone is used as the next candidate, or used as the length of the phrase containing a word. Learning the phrase length that gives the highest priority to the length specified by the person is known.
[0004]
Furthermore, recently, we have focused on specific relationships between words (for example, “hot” and “tea” in “hot tea” or “hot” and “summer” in “hot summer”) and remembered this relationship. By preparing a dictionary, when one word (for example, “tea”) is identified, a word related to the word (for example, “hot”) is selected as the first candidate (for example, JP-A-3-105664 “Kana-Kanji conversion device” and JP-A-4-278761 “Kana-Kanji conversion device”). These specific relationships between words are called “dependency” or “co-occurrence”.
[0005]
[Problems to be solved by the invention]
However, these Kana-Kanji conversion devices only look at the relationship between words in a segment obtained by segmenting a segment once. Memorized dictionaries are also useless. In the first place, since only the relationship between adjacent words is seen as the relationship between words, the most natural kana-kanji mixed sentence in Japanese may not be obtained. However, if you broaden the scope of studying the relationship between words, the number of combinations will increase geometrically as the number of input kana characters increases, thus completing the kana-kanji conversion. It takes an unacceptably long time.
[0006]
In addition, as the relationship between these words, the role of particles in Japanese is overlooked in the information that only two or more words are used in close proximity (hereinafter referred to as “co-occurrence” in a broad sense) A correct conversion result cannot be obtained. For example, just remembering the information that "reporter" and "return to work" have a co-occurrence relationship, even if "quick crazy" can be correctly converted to "reporter will return to work" “To go to work” will be “returned to the reporter” and will not be converted correctly. Therefore, as the dependency information, “noun” + “particle” + “use” (for example, “reporter” + “ga” or “ha” + “go home”), Dependent information including information on adjuncts (particles, etc.) that are allowed to be used can be stored and stored, but in Japanese, the particles that are adjuncts change in expressions such as passive and servants. Therefore, this time invites a problem that “cannot be crisp” cannot be correctly converted. Nonetheless, if the range of allowed adjuncts is expanded in order to allow the formation of dependency in the case of passive or use, the accuracy of kana-kanji conversion by dependency will be reduced.
[0007]
Furthermore, in Japanese, there is a feature that there are some cases where different notation is permitted even for the same word, such as “reception” and “acceptance”, or “acceptance”. In order to accept all of the above, there is a problem that not only all of the written words must be registered in the word dictionary but also in the dependency dictionary. Registering all derivative notations in order to make all these derivative notations subject to dependency not only makes the dependency dictionary very large, The time required also increased, and it was thought that it would be unacceptable.
[0008]
The kana-kanji conversion device and the kana-kanji conversion method of the present invention solve these problems, and use the dependency relationship in consideration of usage and passiveness to split the input character string and change the priority order of the kanji candidates. This was done for the purpose of obtaining a desired kana-kanji mixed sentence, and the following structure was adopted.
[0009]
[Means and Actions for Solving the Problems]
The kana-kanji conversion device of the present invention is
A kana-kanji conversion device that refers to a word dictionary and creates a phrase kanji writing candidate by segmenting an input kana character string and generating a candidate character string constituting a kana-kanji mixed sentence using the phrase kanji writing candidate There,
A dependency information dictionary that stores information on a dependency word and a reception word that constitute a dependency between predetermined phrases together with information on an allowable adjunct allowed between the dependency word and the reception word;
When performing the process of segmenting the input character string, referring to the independent word dictionary storing the independent words and the auxiliary word dictionary storing the auxiliary words, the words between the words that can be generated from the input kana character string Generating means for generating the combination comprehensively;
A phrase search means for searching a phrase having a word corresponding to the information of the dependency word and the received word with reference to the dependency information from the generated combination of words;
First determination means for determining whether a word attached to the related word is the permissible attached word when a phrase having a word corresponding to the related word and the received word information is searched;
When a clause having a word corresponding to the related word and the received word information is searched, if the attached word following the received word is a word representing a use or passive, the word attached to the related word is A second determination means for determining whether the word corresponds to a use or passive character;
When either of the first and second determination means makes an affirmative determination, the combination of the searched words is regarded as a priority candidate in the phrase segmentation, and when neither of them makes an affirmative determination A phrase candidate restriction unit that does not set the combination of the searched words as a priority candidate in the phrase segmentation;
The main point is that
[0010]
The Kana-Kanji conversion method corresponding to this Kana-Kanji conversion device is
A kana-kanji conversion method for referring to a word dictionary, writing an input kana character string into phrases, and generating a kana-kanji mixed character string candidate by a computer,
When performing the process of segmenting the kana character string input from the keyboard, it can be generated from the input kana character string by referring to the independent word dictionary storing the independent words and the auxiliary word dictionary storing the auxiliary words A comprehensive combination of words is generated by a computer,
Dependency information is stored by referring to the dependency information stored in the combination of the generated words together with the information on the allowable adjunct allowed between the dependency words. Use a computer to search for phrases with words that correspond to the received information,
When a phrase having a word corresponding to the related word and received word information is searched and found,
Determining whether the word attached to the dependency word is the allowable adjunct;
If the adjunct word following the received word is a word representing a use or passive character, it is determined whether the word attached to the related word is a word corresponding to the use character or passive;
If any one of the determination results is an affirmative determination, the combination of the searched words is set as a preferential candidate in the phrase segmentation, and if any determination result is not an affirmative determination, Limiting the phrase segmentation candidates by a computer by not using a combination of searched words as a priority candidate in the segment segmentation.
Is the gist.
[0011]
According to the first kana-kanji conversion apparatus and the kana-kanji conversion method of the present invention configured as described above, both the dependency words and the received word information constituting the dependency between predetermined phrases are stored in the dependency information dictionary. This is stored together with information on allowable adjunct words allowed between words, and when the input character string is segmented, the dependency information stored in the dependency information dictionary is referred to. Search for phrases with words that correspond to the information. When a clause corresponding to the related word and the received word information is searched, it is determined whether the word attached to the related word is an allowable adjunct, and the adjoining word following the received word is a word representing a working or passive word. Determines whether the word attached to the dependency word is a word corresponding to the use or passive. Based on the search result, the phrase segmentation candidates are limited. Therefore, when there is dependency information, instead of simply looking at the establishment of the dependency due to the presence of both words, it is assumed that the dependency attached when the word attached to the dependency language allows the relationship between the two words, If the received word is a word indicating a service or passive, and if the word attached to the dependency word is a word corresponding to the service or passive, the phrase candidate is limited as the formation of the dependency. Undesired candidates for the split writing are difficult to select, and there is a high possibility that the desired split writing is performed. If the priority order of kanji candidates for each phrase is to be changed on the premise of phrase segmentation that has already been estimated by another method instead of the restriction of phrase candidates, the desired kanji candidate is set as the first candidate. The possibility of being obtained can be increased.
[0012]
Here, the dictionary is a word dictionary storing a word reading and a notation corresponding to the reading, and when there are a plurality of notations as a notation corresponding to the word reading, it is determined as a representative notation. And a notation information section that stores the notation defined as a derived notation,
The phrase search means is a means for searching for the dependency using only the representative notation stored in the word dictionary,
Furthermore,
The phrase segmentation candidate may include candidate character string display means for displaying the converted candidate character string using the representative notation and the derived notation stored in the notation information part of the word dictionary.
[0013]
In this kana-kanji conversion device, a representative notation and a derived notation are stored in the word dictionary, and only the representative notation is used for the dependency test for each phrase written in the first kana-kanji conversion device. A word is specified by using and a candidate character string for the specified word is displayed using a representative notation and a derived notation.
[0014]
Furthermore, in the kana-kanji conversion device,
Instead of the phrase candidate restricting means, a kanji candidate priority means for changing the priority order of kanji candidates for each phrase based on the determination results of the first and second determining means may be provided. In this case, although those that do not fall under dependency are left as candidates for kanji, those having dependency are preferred as candidates for kanji.
[0015]
Moreover, the phrase search means in the kana-kanji conversion device is
A backward search means for searching for phrases including words corresponding to the dependency information in order forward, excluding the already registered searched range, starting from the backward phrase;
Searched range registration means for registering from the starting phrase to the found phrase as a searched range of dependency information when a phrase having a word corresponding to the dependency information is found by the search Can also be provided. In this way, at the time of the next search, this range is excluded from the search range, so that the search time for the division can be shortened, and there is no possibility of erroneously selecting linked dependency.
[0016]
The phrase candidate restriction means in the kana-kanji conversion device is preferentially selected from the phrase segmentation including the word when the first or second determination means determines the presence of the corresponding quasi-word, the received word, and its associated word. If a means for selecting is provided, it is preferable to preferentially use the split writing in which the dependency is established.
[0017]
In the kana-kanji conversion device comprising the kanji candidate priority means,
This kanji candidate priority means, when the first or second determination means determines the presence of the relevant vocabulary, receiver, and its adjunct, preferentially selects the phrase segmentation including the word, Means for selecting the word as a first candidate for kana-kanji conversion may be provided. In this case, the division and the selection of the first candidate are made based on the dependency, and the establishment of the dependency has the highest priority.
[0018]
Further, the phrase search means includes means for determining that the dependency relation is established when the auxiliary word existing between the words having the dependency relation is a word having a predetermined specific grammatical structure. If it is provided, it is possible to reduce the amount of information related to attached words.
[0019]
In the kana-kanji conversion device provided with the phrase search means,
The phrase search means
Starting from a predetermined phrase, refer to the dependency information stored in the dependency dictionary, and search for a phrase having a word corresponding to the dependency information up to a phrase other than the phrase adjacent to the phrase as the origin. Means to search for alternate phrases,
Search range exclusion means for excluding the range from the starting phrase to the found phrase from the search range by the next phrase search means when a dependency relationship is found by the separate phrase search means;
It can also be provided.
[0020]
In this case, it is possible to avoid the erroneous determination of the correct dependency by eliminating the intersection of the dependency and to speed up the search for the dependency.
[0021]
In such a kana-kanji conversion device, a known phrase segmentation technique can be applied to a range in which a phrase including a word having a dependency relationship is not found with reference to the dependency dictionary. For example, a two-segment longest match method may be used, and a score (cost) is added to the likelihood of coupling between words or / and between phrases, so that the likelihood of coupling is maximized (minimum cost). A word or / and a phrase may be selected. Desirably, a combination that maximizes the likelihood of coupling between words and between phrases may be selected.
[0024]
In this kana-kanji conversion device and kana-kanji conversion method, when performing dependency testing with reference to the dependency information dictionary, the test is performed using only the representative notation registered in the word dictionary, and the input character string is converted. Although the word which comprises is specified, when displaying the candidate character string of this specified word, not only the representative notation memorize | stored in the word dictionary but a derivative notation is displayed. Therefore, it is possible to achieve both high-speed grammar processing including dependency and diversity of notation.
[0025]
【Example】
In order to further clarify the configuration and operation of the present invention described above, preferred embodiments of the present invention will be described below. FIG. 1 is a block diagram showing a control logic for kana-kanji conversion, and FIG. 2 is a block diagram showing hardware on which the kana-kanji conversion control logic actually operates. As shown in FIG. 2, this apparatus includes the following units connected to each other by a bus 31 around a known CPU 21. The units connected to each other by the CPU 21 and the bus 31 will be briefly described.
[0026]
ROM 22: mask memory for storing a kana-kanji conversion program, etc.
RAM 23: Readable and writable memory constituting main memory,
Keyboard interface 25: An interface for managing key inputs from the keyboard 24.
CRTC 27: CRT controller that controls signal output to the CRT 26 that can display in color.
Printer interface 29: an interface for controlling output of data to the printer 28;
A hard disk controller (HDC) 30; an interface for controlling the hard disk 32;
It is. The hard disk 32 stores various programs loaded into the RAM 23 and executed, a kana-kanji conversion processing program provided in the form of a device driver, or various conversion dictionaries referenced by the kana-kanji conversion processing program.
[0027]
With the hardware configured in this manner, text can be input, kana-kanji conversion, editing, display, printing, and the like. That is, the character string input from the keyboard 24 is subjected to predetermined processing by the CPU 21, stored in a predetermined area of the RAM 23, and displayed on the screen of the CRT 26 via the CRTC 27.
[0028]
Next, functions executed by the hardware thus configured will be described with reference to FIG. The configuration and operation of each unit shown in FIG. 1 will be outlined. The processing performed here is executed by the central processing unit (CPU 21) based on data input from the keyboard 24. All processing is performed by the CPU 21. As for kana-kanji conversion, a predetermined interrupt process is activated when the keyboard 24 is operated, and the input key image is converted into a corresponding kana character string, and further converted into a kana-kanji mixed character string. Starts. Of course, in a computer capable of parallel processing, kana-kanji conversion may be performed by one application (input method), and the conversion result may be transferred to a required application. In this case, inputs from the keyboard 24 are collectively accepted by the input method.
[0029]
The key image from the keyboard 24 is received by the character input unit 40 and is converted into a corresponding kana character string here. In the case of romaji input, it is converted into a kana character string with reference to a predetermined conversion table. Each time one kana character is obtained, the character input unit 40 sends the kana character to the conversion control unit 42. The conversion control unit 42 plays a central role in kana-kanji conversion, controls various kana-kanji conversion described later, and sends the result to the converted character string output unit 44. The post-conversion character string output unit 44 actually sends a signal to the CRTC 27 and displays the post-conversion character string on the CRT 26.
[0030]
The conversion control unit 42 passes the received kana character to the character string input unit 50. The character string input unit 50 stores the kana character string in the character storage unit 52. Based on this character string, the independent word candidate creation unit 54 and the adjunct word candidate creation unit 64 create word data candidates. The independent word candidate creation unit 54 uses the independent word dictionary 58 stored in advance in the hard disk 32 and performs processing for extracting independent word candidates from the obtained kana character string under the management of the independent word analysis position management unit 56. Do. On the other hand, the adjunct word candidate creation unit 64 uses the adjunct word dictionary 68 and performs processing for extracting an adjunct word candidate from the obtained kana character string under the management of the adjunct analysis position management unit 66. The process of extracting independent word candidates and auxiliary word candidates while moving the analysis position will be described later.
[0031]
Here, the independent word dictionary 58 changes priorities such as homonyms and affixes by learning. This learning process is performed by the dependency learning unit 70, the independent word learning unit 72, the auxiliary word learning unit 74, the affix learning unit 76, and the character conversion learning unit 78. The dependency learning unit 70 gives priority to the combination selected by the user in the same word combination when the user selects a word other than the word corresponding to the dependency under the condition that the dependency is satisfied. It learns dependency relationships. The independent word learning unit 72 learns the last selected word as the highest priority candidate in the independent word group in which the homonyms exist. The auxiliary word learning unit 74 learns which word form is used to convert an auxiliary word such as “please”, for example, “please” or “please”. Further, the affix learning unit 76 learns conversion formats (for example, “Go”, “Go”, etc.) such as prefixes and suffixes. The character conversion learning unit 78 learns a character string when the input character string is confirmed as hiragana or katakana as it is, and outputs the hiragana or katakana determined in the subsequent conversion processing as a candidate. .
[0032]
The independent word candidate creation unit 54 and the attached word candidate creation unit 64 obtain the created word candidates, and the word data creation unit 80 creates data for each word candidate. That is, as a result of the connection verification unit 82 referring to the connection verification table 84 for connection between the obtained independent words and ancillary words, independent words and independent words, and connection between phrases consisting of “independent words + ancillary words”, The result of the cost calculation unit 86 performing the overall cost calculation is obtained and output as data for each word. This word data is temporarily stored in the word data storage unit 100, receives the adjustment output from the dependency candidate adjustment unit 90, and is used for the phrase segmentation process.
[0033]
The dependency candidate adjustment unit 90 receives the word candidates from the independent word candidate creation unit 54 and the adjunct word candidate creation unit 64, and tests the dependency. The dependency test is performed by referring to a dependency dictionary 98 prepared in advance on the hard disk 32. The range in which the dependency test is performed is managed by the dependency range management unit 96. In addition, there are several allowable conditions for the dependency relationship test, which are determined by the service / passive analysis unit 92, the particle allowable analysis unit 94, and the like. The dependency candidates adjusted by the above-described dependency test and the word data described above are integrated by the word data storage unit 100 and used for the phrase segmentation processing by the phrase segmentation unit 102. The phrase segmentation unit 102 determines a first phrase segmentation candidate from the obtained data.
[0034]
With the above processing, the first candidate for phrase segmentation and the first candidate for kana-kanji conversion for each phrase are determined. The phrase segmentation unit 102 stores the candidate in the phrase data storage unit 106, and the stored candidate is output to the conversion control unit 42 by the conversion character string output unit 108. The conversion control unit 42 displays this character string as a candidate character string, and an undesired character string may be a candidate. Therefore, in response to an instruction from the user, processing such as display and selection of the next candidate is performed. Do. These instructions and selection results are input to the phrase data storage unit 106 and each of the learning units 70 to 78 described above, and are used to determine part of the phrase or rewrite priority by learning. Although not shown, when the character string is finalized by the user, all data temporarily stored in each unit is deleted and prepared for the next conversion.
[0035]
The outline from the input of the kana character to the output of the conversion word character string has been outlined above. Next, details of each process will be described. First, general phrase segmentation processing will be described, and then dependency processing, which is a main part of the present invention, will be described. FIG. 3 is a flowchart showing an outline of phrase segmentation processing by the minimum cost method. As shown in the figure, first, after performing initialization processing (step S200) such as erasing temporarily stored data and initializing the analysis position to the first digit, processing for obtaining the analysis position is performed (step S200). Step S210). The analysis position is a position where the kana character string input so far is advanced one by one in order. For example, if the kana character string “Kagaku wa hako hakubu” is input as shown in FIG. 4, the first analysis position is the position of the first digit “ku”. At this analysis position, a process for searching the independent word dictionary 58 and the attached word dictionary 68 stored in the hard disk 32 is performed (step S220).
[0036]
After searching the dictionary, a process for checking the combination of the obtained word with the previous word is performed (step S230). If only words that cannot be combined between words are obtained, the dictionary is further stored. Search for. For example, in the example shown in FIG. 4, “wa” of the counsel retrieved from the adjunct dictionary 68 for “ha” of “kohahakobu” cannot be combined with the immediately preceding case particle “wo”. Therefore, it is treated as invalid data by the connection verification by the word data creation unit 80 and the connection verification unit 82. In FIG. 4, a symbol “x” is attached to a word that is determined to be invalid by such a combination check. The connection between words is stored in advance in the connection verification table 84. This connection verification table 84 is a table that gives information on the possibility of connection between parts of speech of a word. It is given as a matrix of × 400. When the dictionary search and the combination check at one analysis position are completed, the analysis position is advanced in order and the process is further repeated.
[0037]
For words that are likely to be combined, cost calculation is performed next, and processing for obtaining the minimum total cost of the words is performed (step S240). This process is performed by the cost calculation unit 86. In the example shown in FIG. 4A, “car” is, for example, “ku” + “ru” + “ma”, “car” + “ma”, “car”. , And when applying words to them, it is assumed that it has the cost of independent words = 2 and attached words = 0, and if it is “suffer” (independent words) + “style” (independent words), The total cost of “flow” is calculated as 4. At this time, the cost of “between” is 4 because the minimum total cost is calculated, and “come” + “between” is not the cost 6 of “bitter” + “flow” + “between”. This is because the case cost 4 is adopted. Since “de” and “ha” are attached words, the cost of the word “car” = 2 with the lowest cost among the previous words becomes its own cost. In FIG. 4, the cost of each word is shown in the lower right.
[0038]
After the above cost calculation, the cost of each word is checked, and a process of invalidating an inappropriate cost is performed (step S250). Inappropriate cost is a combination of words that is costly compared to other word combinations. In other words, selecting a combination of words such as “ku” + “toru” will be more expensive than other words “coming” and “repeating” obtained up to that position. Judgment is made and excluded from the phrase candidates. In FIG. 4, words that are not adopted based on this concept of minimum cost are shown as “●” in the upper right of the word. In FIG. 4, “◯” indicates that the word remains as a word that may form a phrase candidate as a result of the above-described combination check and cost check.
[0039]
Next, a process for linking the word candidates thus given the cost is performed (step S260). That is, for words for which coupling is valid, the coupling relationship is related by setting a pointer. In the example of FIG. 4, “come”, “carry”, “car”, “to”, “de”, “ha”, “ha”, etc. have been calculated as the minimum total cost. It links to “to” and “car” to “de” and “to”. This combination check, cost calculation, and linking process are repeated while searching for all words at one analysis position is completed. Further, when the dictionary search at the analysis position is completed, the analysis position is further advanced by one, the establishment of a new word is examined, and the connection check and cost calculation are repeated in the same manner.
[0040]
When the analysis position reaches the position of the last input kana character and the analysis is completed for all words (step S265), a process of searching for the path with the minimum cost is performed on the assumption of the above processing (step S265). Step S270). This process is performed by the phrase segmentation unit 102, and is a process of searching for a combination of valid words that minimizes the sum of the costs assigned to the words. In the example of “Humps until coming”, as shown by the solid line J in FIG. 4B, the division cost of “by car” + “carrying a box” + “carrying” is a total cost of 18, so the minimum Selected as a cost. Although not the minimum cost, other phrase segmentation candidates are also searched. For example, as shown by a broken line B in FIG. 4B, the division is “by car” + “child” + “carry” (cost = 20). In this way, after the candidate for the division is created (step S280), the process for creating the candidate inside each phrase is performed (step S290). That is, for example, candidates such as “box” and “box” are prepared for “hako wa” within one phrase segment. These phrase candidates and word candidates are used when the user is instructed to change the way of segmenting or to display the next candidate.
[0041]
An example of another phrase segmentation is shown in FIG. This example is used to explain the dependency described later, but shows a case where the phrase segmentation by the minimum cost method that does not consider the dependency is applied to “make a person squeak”. In this example, the combination check (step S230) excludes “ゃ” because the direct combination with the noun is invalid (x mark). Also, “Thu”, “Mr.”, “Company”, etc. are excluded from the minimum total cost check (step S250) (marked with ●). As a result, in the example shown in FIG. 5, the phrase segment “Kisakuo” + “Kisaku” is written, and if the priority order of each word is the order shown in FIG. “Your company” will be selected as the first candidate. As for “Kisha” in the second half, if the analysis has only reached the end of “Kisha”, the same “Your company” as the first candidate in the first half will be selected. When “seed”, “do”, etc. are attached after “”, “returning to office” can be displayed as the first candidate in consideration of, for example, the word “usage” attached.
[0042]
In the above explanation, the cost calculation is performed only for each word itself. However, in practice, a point for reducing the cost is given according to the degree of ease with which words are joined, or clauses are joined. Similarly, it is possible to give a point for reducing the cost to a combination of phrases that are easily combined based on grammatical rules. Here, in order to facilitate the understanding of the phrase segmentation process, only the simplest method has been described.
[0043]
Based on the above phrase segmentation process, the phrase segmentation process by dependency will be described next. FIG. 6 is a flowchart showing a process for performing a dependency test. This process is performed in parallel with the processes of steps S220 to S250 shown in FIG. Actually, after obtaining the analysis position (step S210), when searching various dictionaries, the dependency dictionary 98 is also searched along with the search of the independent word dictionary 58 and the attached word dictionary 68, and the combined check and the minimum total The following dependency verification process is performed together with the cost check accompanying the cost calculation. When this process is started, first, the dependency candidate adjustment unit 90 searches the dependency dictionary 98 for words that can be candidates at the analysis position (words with a circle), and whether the dependency information exists is a word. It is determined whether or not (step S300). Note that if the word at the analysis position is a conjunction, a moving verb, or an independent word, there is no dependency, and the processing for that word is immediately terminated.
[0044]
As an example, a case will be described in which the kana character “Kisaku Seki” is input and the analysis proceeds to “Kisaku Sakusha”. At this time, “reporter”, “your company”, “train”, “return home”, etc. are obtained as candidates for “Kisha” in the latter half, so whether there is any information in the dependency dictionary 98 for each of these words Is examined. If the word to be received is a predicate such as “heard” or “handed” and has a usage form, the word “hearing” or “handy” or the basic form “listening” or “handed” is used as a keyword. Thus, the dependency dictionary 98 is configured to be referable.
[0045]
An example of the structure of the dependency dictionary 98 in the embodiment is shown in FIG. The dependency dictionary 98 of the present embodiment has a structure in which [headline + reception language (stem) +1 dependency word + attached word information] as a unit. In the example of FIG. Headline “Kissha” + Receiving Word “Return to Home” + Spoken Word “Reporter” + “Ga”, Headline “Kisha” + Spoken Word “Returning Home” + Spoken Word “Your Company” + “Ni”, Headline “Kissha” + There are multiple sets of data for a single spoken language, such as the spoken language “returning to the office” + clerk “train” + “de”. Furthermore, for the headline “Kissha” and the received word “Reporter”, the headline “Kissha” + the received word “Reporter” is similarly applied to each word constituting the related word group “Your company, newspaper, news agency ...”. ”+ Linguistic words“ your company ”+“ no ”etc., each linguistic word has data. These data are arranged in the order of the Japanese syllabary of the headwords for the received words. Of course, similar dependency information is stored for other words. The dependency candidate adjustment unit 90 searches the dependency dictionary, and when there are corresponding headlines and received words, the candidate candidates are taken out of the dictionary and used for the dependency test. Since these data have the same heading and spoken word, the heading word and the catching word may be prepared at the head of the entire data group, and the information of the related words and the attached words may be individually prepared.
[0046]
Further, in the example of FIG. 7, only the minimum information is shown in order to show the dependency test in an easy-to-understand manner, but the actual dependency dictionary 98 includes “received word header + dependent word header”, “received word header length”. ”,“ Received Word Kanji ”,“ Received Word Part of Speech ”,“ Dependency Word Heading Length ”,“ Modified Word Kanji ”,“ Dependent Word Part of Speech ”,“ Dependency Relationship ”, and the like. The part-of-speech of the received word or the dependency word is necessary for examining the establishment of the dependency and the admissibility of the attached word, and the heading length data is necessary for cutting out from the dictionary 98 at high speed.
[0047]
If a word for which dependency information exists (hereinafter referred to as a received word) is found, the next word corresponding to the dependency is forwarded, except for the range where the dependency has already been established. A search is performed to determine whether or not there is a related term (step S310), and it is determined whether or not there is a corresponding related term (step S320). At this time, the search for the related words is performed not only for the word having the minimum total cost but also for other words. Now, in the dependency dictionary 98, as shown in FIG. 7, “reporter (returned to the company)”, “returned to the company”, “customer (to) reporter”, and “return to the train (train)” It is assumed that the receiver is stored. Here, the kana in () is an adjunct that is allowed as possibly existing between words that have a dependency relationship. For the “return to office” and “reporter” determined to be possible to be received in step S300, there are received words in the dependency information. When it is determined whether there is a related word for each of them, in the example shown in FIG. For "Return to work", the character string candidates "Kisha" are judged as "Reporter", "Your company", and "Car". For "Reporter", the character string candidate "Kisha" is "Your company" Is determined to be applicable. If it is determined in step S320 that there is a dependency word, it is next determined whether or not an adjunct word existing between both words is a word that allows the presence of dependency (step S330). .
[0048]
The particle admissibility analysis is to determine whether or not the allowable relationship defined by the dependency type is viewed, and has the following types for each dependency type.
[I] Continuous modification type
▲ 1 Noun + particle + particle in the case of a predicate
Case particles "ga""from""de""to""ni""to""more""to""no"
Particle particle "ha"
(2) In the case of idioms + idioms
(3) Optional particle in the case of noun + predicate (particle abbreviation type)
"GA""HA" particle, adjunct particle
[II] Complex modification type
(4) Noun + particle + particle in case of noun
"of"
(5) Particles in the case of body + body (parallel)
"Ya""To"
(6) In the case of predicate form + noun
(7) In case of conjunction + noun
[0049]
That is, it is assumed that the relationship between two words determined to be in a dependency relationship belongs to any one of the above-mentioned (1) to (7), and an adjunct word existing between both words in the dependency relationship (mostly Is a particle or a particle expression), the dependency dictionary 98 is set with an allowable particle for a word having a dependency relationship, and this is verified. For example, when the dependency between “promotion” and “handedness” is accompanied by the allowable setting of the particle (no ·), it belongs to the case (1) above (noun + particle + probe). “” And “ga” can exist between the two words (smooth, savvy → ○), but other case particles “kara” and “de” are unacceptable. Would be good →→).
[0050]
Regarding the relations (1) to (7), it is judged that the relations other than those listed are allowed. Examples of determinations that are permitted are listed below, but these may include cases where the dependency does not hold as an actual expression. However, dependency is a broad concept as an actual human language activity, and a too strict dependency arrangement is often not suitable for reality. Also, a too strict dependency arrangement will only lead to a mischievous increase in the dependency dictionary 98 and will also reduce the speed of dependency verification. Therefore, in this embodiment, with respect to the admissibility of attached words, the relations in which dependency occurs are divided from (1) to (7), and those that are clearly permitted or not permitted are permitted in the dependency dictionary. It is memorized as a word with a dependency relationship, and the others are allowed.
[0051]
[III] Permissible expressions-in the case of continuous modification
・ Noun + case particle expression + case particle expression in predicate
“By”, “As”, “For”, “In”, “By”, etc.
・ Noun + particle + verb particle
“Some”, “Even”, “Shi”, “But”, “M”, etc.
・ Noun + adverb + adjunct in adjunct
"Kiri""About""One by one""Only", etc.
・ Noun + adverbial expression + adverbial expression in predicate
"So,""If", etc.
・ Phrase + particle + particle
"Noha" etc.
・ Connecting particles “So” “From” “From” “Te”, etc.
・ Conjunctive particle-like expressions
・ Expression that parallels preaching + preaching “ka” “shi” “ri” “simultaneously”, etc.
[0052]
[IV] Permissible expressions-complex modifications
・ Noun + particle-like expression + particle-like expression in noun
"In", "related", "based on", etc.
・ Phrase + particle-like expression + particle-like expression in noun
"For", "like", "with", "etc."
・ Expression that parallels body language + body language "ka".
[0053]
According to the above rules, the admissibility of an adjunct word between two words for which a dependency relationship is found is determined. In the case of “reporter” and “return home” as an example, the allowable case particle is “ga”, and therefore, the establishment of dependency is not permitted for “Kisha wo Kisa”. Therefore, this is judged (step S340), and if it is determined that the dependency does not hold despite the presence of the dependency word and the reception word to establish the dependency, the next use / passive relationship is determined. A receiving test process is performed (step S342).
[0054]
The use / passive dependency test process is performed by the use / passive analysis unit 92 shown in FIG. This process will be described in detail. In the sentence example shown in FIG. 5, when the analysis further proceeds to “quickly crawl” as shown in FIG. 8, it can be determined that the service is active / passive, and the dependency processing considering the case of active / passive is performed. This is to be performed (step S342). Similar to steps S320 and S330, this process is a determination as to whether there is a corresponding related word and whether an attached word in that case is an allowable word. Focusing on “return to work” for “Kisha”, the corresponding word “reporter” exists, and then the analysis of the ancillary word shows that in the case of a service, the original adjunct “ga” is “ Since it is previously stored that “is allowed”, it is determined that the dependency is established. The relationship between “returning to work” and “your company” is “Ni”, which is an allowed adjunct. It is not allowed. Similarly, “returning to office” and “train” + “de” are also tested for dependency, and it can be seen that there is no admissible adjunct between the dependency and the reception.
[0055]
Therefore, it is determined whether the dependency is established based on the analysis results (step S344). If it is determined that the dependency is established in consideration of the use / passive, the normal dependency is determined in step S340. When it is determined that is established, the phrase including the word is preferentially processed as a phrase candidate beyond the difference in the minimum total cost (step S350). Further, the range from the received word to the related word thus found is registered as a dependency established range, a process for managing this is performed (step S360), and it is determined whether or not dependency search has been performed for the entire range ( Proceed to step S370). In addition, when the establishment of the dependency is denied in consideration of the use / passive as well as the normal dependency, the process proceeds to step S370 without performing steps S350 and 360.
[0056]
The above processing (steps S310 to 370) is repeated until the dependency is searched forward from the received word and all words except the range registered as searched have been completed. When completed, it is next determined whether or not the dependency test has been completed for a plurality of candidates for the received language (step S380). In other words, in this example, the candidate “your company”, “return home”, “reporter”, “train”, etc. for the latter half of the spoken word “Kissha” is tested to see if there is any dependency relationship. It is. A word for which a dependency relationship is established is found, the admissible analysis of the attached word is passed, and a word that is determined to have a dependency relationship is set to the highest priority as a phrase candidate (step S350). When a dependency relationship is established for a plurality of candidate words, phrase candidates with higher priorities are set in the order registered in the dictionary.
[0057]
Here, a description will be given of the fact that when a phrase including a word with a dependency is set as a phrase candidate, the phrase is prioritized even if the phrase includes a word whose minimum total cost is not reached. In the example of “Kisakusakushaseki”, the selected phrases “Kisakuo” and “Kisashise” are phrase candidates based on the minimum total cost method that does not perform dependency tests, and how to divide the phrases themselves. The same. However, for example, assuming that the kana character string “Jikkishashashasha” is input and the independent word “steam train” exists, as shown in FIG. "Becomes the first candidate with a minimum cost path (solid line G). On the other hand, if the dependency relationship (“Reporter (G)) Return to work”) is tested by the working / passive, “the above”, “Reporter” and “Return to work” will not be the minimum cost path. One candidate is selected (broken line B in FIG. 9).
[0058]
FIG. 10 shows phrase candidates that are finally obtained by deleting candidates that cannot be connected by performing connection check between words or the like for the first example sentence “Make a chat”. Therefore, the first candidate for the phrase segmentation is “return the reporter to the office”.
[0059]
As described above, an example of the usage expression has been described. As types, there are the following as nouns N1 and verbs P.
“N1 + P” is “N1 is + P”.
“N1 + P / Can” is “N1 + P”
“N1 + P make / can do” is “N1 + P”
[0060]
In the present embodiment, the following types of usage that require particle verification over two or more phrases are not allowed.
“Make + P (verb form) for N1 +”
“+ N1 to + P (adjective verb)”
"+ N1 + P (adjective)"
“+ N1 to + P (noun)”
“+ N1 like + P (verb ending form)”
“Make N1 + P”
[0061]
Furthermore, the dependency process in the case of passive is illustrated. An explanation will be given by taking as an example a process in which the input kana character string “I can tell you” is written as a phrase when “student (teaching)” exists as a dependency. FIG. 11 shows a case in which the analysis position advances to “Tell me a word” and “Teach” is searched backward to find whether there is a related word. There are not only "teaching" and "teaching" as well as "teaching" but also "teaching mathematics" and "teacher". These are the dependency dictionary 98. And can search for “teaching” as a headline. Since this search is performed retroactively, the test starts from “O” and “Tail”, then “Saint is” and “Student is” and “Student ( You will find “students” in “Teach”. Since this dependency is “student teaches” as it is, the acceptance analysis of the attached word is not passed, and the dependency test is temporarily terminated. After that, when the verification of the phrase segmentation progresses to “assertion” and the dependency test is performed again, the admissible analysis of the attached word is passed for the first time by the test of the active / passive dependency test. Therefore, it is determined that it is established as a dependency in the case of passive, and “Student” and “Teached” are the first candidates for phrase segmentation. This is shown in FIG. The obtained first candidate is “student taught”. Here, the range from “student” to “taught” is excluded from the subsequent search range of dependency as a dependency formation range.
[0062]
Although an example of the passive expression has been described above, examples of the type include those described below as the noun N1 and the verb P.
“N1 is + P is / can be” but “N1 is + P” passive type
"N1 is + P / can be" is "N1 is + P" passive type
"N1 + P can / can't" is "N1 + P" passive type
[0063]
In the present embodiment, the following passive types that require particle test over two or more phrases are not allowed.
“N1 + P)”
“N1 + P)”
“N1 is + Ped to + N2”
“N1's + P”
“N1 is + P (sa) to + N2”
“+ N2 from + N1”
[0064]
According to the present embodiment described above, since the dependency cost information including the case of active / passive is simultaneously searched in the process of calculating the word cost and obtaining the phrase segmentation candidate, the phrase segmentation candidate is selected. At the requesting stage, it is possible to reflect the dependency information including duty / passive. Since the dependency information is a high-level language activity, it avoids adverse effects due to the uniformity of the selection of segmentation by the cost calculation between words and between clauses, and the candidate for segmentation according to the user's intention It can be obtained. In addition, since the dependency processing is performed simultaneously with the phrase segmentation processing by the minimum cost method performed with reference to the independent word dictionary 58 and the attached word dictionary 68, the phrase segmentation processing using the dependency information is performed in a short time. Can be completed within. In the case of referring to the dependency dictionary 98 again after completing the phrase segmentation, not only the method of dividing the phrase cannot be changed using the dependency information but also the reference of the dictionary is performed again. It takes time.
[0065]
In addition, if it is determined that the dependency has been established, the range from the received word to the dependency word is excluded from the subsequent search range as the dependency establishment range, so the dependency ranges may intersect. Absent. In addition, it is not determined that two or more spoken words receive a single dependent word. In addition, since the establishment of the dependency is judged beyond the adjacent clauses, the dependency test can be correctly performed even if the modification by the adverb or the like enters between the dependency relationships. Therefore, when a plurality of dependencies are established, as shown in FIG. 13A, a combination in which independent dependencies are established separately, or as shown in FIG. Or a combination in which another dependency is established so as to straddle one dependency, as shown in FIG. 13C.
[0066]
Next, a second embodiment of the present invention will be described. In the second embodiment, the same hardware configuration as that of the first embodiment is used, and the functional blocks thereof are almost the same as those shown in FIG. The functional blocks differ in the structure of the independent word dictionary 58 and the structure of the dependency dictionary 98, which are a word search process, a dependency test process, and a display process associated with the difference in the dictionary structure. These differences and the differences in the dictionary structure will be described in order according to the processing in the second embodiment.
[0067]
FIG. 14 is a flowchart showing a kana-kanji conversion processing routine in the second embodiment. This processing routine is started when a conversion key (for example, “space key”) is pressed after one or more kana characters are input from the keyboard 24. When a predetermined number of kana characters are input even when the conversion key is not operated, or when a delimiter such as “.”, “,” “.”, Or the like is input, the kana-kanji conversion processing of FIG. Can be started. When this process is started, first a word search process (step S400) and a segmentation process (step S420) are performed. These processes correspond to the processes of FIG. 3 in the first embodiment.
[0068]
FIG. 15 shows details of the word search processing routine. As shown in the figure, when the word search processing routine is started, first, a process of setting the word search start position M to the value 1, that is, the start position of the input kana character string is performed (step S402). Next, the variable L indicating the reading length in the word search is initialized to the value 1 (step S404), and the word having the reading length L is searched from the independent word dictionary 58 and the attached word dictionary 68. Processing is performed (step S405). Here, as shown in FIG. 16, the independent word dictionary 58 includes a header, an index, and a dictionary body. The header is information for managing the dictionary itself. The index and the dictionary body are managed separately for basic words, derived words, and semantic examples. A basic word is defined in advance as a word representing these notations when one word has a derived notation, for example, when “handling” or “handling” is allowed as a notation for “handling”. Is the word. In other words, the basic word only means a word that is typically used in phrase segmentation and dependency processing. Since it is a word recorded in the word dictionary, it is called a representative word, but it does not basically mean a word in the language. Hereinafter, the basic word is referred to as “representative notation” for display, and the derived word is referred to as “derivation notation”.
[0069]
The area for the semantic example is an area in which the same information as the information regarding the dependency described in the first embodiment is managed. Therefore, in the second embodiment, the independent word dictionary 58 and the dependency dictionary 98 are integrated. The information stored in the semantic example area is dependency information centered on the basic word. The relationship between dependency information and basic words and derived words will be described later.
[0070]
These basic words, derived words, and semantic examples are managed by the B-Tree structure in the dictionary body. The B-Tree structure is a well-known management structure that is adopted when searching for a large number of data. If there are a large number of data, if the data is properly organized, the process until the target data is reached. It is known as a structure in which time becomes an average time. An example of the B-Tree structure in the dictionary body is shown in FIG. When the B-Tree control block is traced to the word block based on the reading (kana character string), the actual word data is placed here.
[0071]
The word data such as the basic word region roughly has a data structure shown in FIG. That is, there is data indicating the data length X of the word data at the head, followed by actual data of X bytes. The headword length Y is recorded at the head of the actual data, and then a Y-byte headword is recorded. The actual word data follows. The word data has a word length W recorded at the beginning thereof, and a 1-byte flag indicating the presence or absence of kanji data is recorded immediately after the word data. After the flag, kanji data is recorded. This kanji data is composed of a kanji code length indicating a kanji data length and an actual kanji character string. Thereafter, word information and part-of-speech information (in some cases, a plurality of part-of-speech information) are recorded. The word information includes data indicating the length of the word information and actual word information.
[0072]
As described above, information on a desired word can be extracted based on the heading character string of the word using the B-Tree structure for both the basic word and the derived word. As shown in FIG. 19, these pieces of word information are composed of a separator and subsequent data. As separators, a display kanji separator that indicates that the subsequent data is displayed kanji data, a reading information separator that indicates that the data is reading information, a derivative notation separator that indicates that the data is derived notation, etc. There is. The display kanji is a record in which kanji corresponding to the representative notation is recorded in order to display kanji by default when one word has a representative notation and a derived notation. In the derivation notation separator, as shown in FIG. 19, the lower 3 bits of the separator correspond to the number of derivation notations, and the subsequent derivation notation 1 and derivation notation 2 are representative notations as illustrated in FIG. The form of transformation with respect to is indicated by a number. That is, if the derivation notation 1 is, for example, the number 5, if the representative notation is “memory”, it means that “with long sound” exists as a derivation notation, and means “memory”. That is, as the information of the derivation notation, the actual derivation notation itself is not stored, but the form of the derivation notation is stored as a number. The reading information gives a reading of the word, and gives the reading when the headword is a Chinese character. This information is used at the time of associative conversion or the like for searching for another kanji having the same meaning from the kanji. Note that one basic word and the derived word (derived notation) corresponding to this basic word (representative notation) are managed in different areas, but the word reading (heading) matches the word part-of-speech information. It is considered that there is a corresponding relationship for what to do.
[0073]
Returning to FIG. 15, after performing a dictionary search (step S405), it is determined whether or not a word matching the kana character string of the length L is found from the start position M (step S406). If the corresponding word is found, next, a process for reading data related to the derivative notation among the data attached to the representative word is performed (step S408). If there are various derivation expressions for the representative word and the derivation word is stored in the derivation word management area, a flag indicating whether or not replacement with the derivation expression is necessary is stored in the attached data of the representative word. ing. Therefore, it is determined whether or not the flag is referred to and the instruction is made to replace the representative word with the derived word (step S410). If this flag is set to a value indicating that replacement with a derived word is required, the representative word searched earlier is expanded in the expansion buffer, and a mark is added to the word (step S411). The expansion buffer is a storage area for expanding all representative words and adjunct words that can form the kana character string for the input kana character string, and is a memory area secured on the RAM 23. .
[0074]
If the instruction to replace the derived word is not given, or the replacement to the derived word is instructed and a mark is given to the representative word, the process returns to step S405, and the word with the reading length L is returned. Repeat the search process. If it is determined that the word with the reading length L no longer exists in the free word dictionary 58 (step S406), it is determined whether or not the length of the search word can be increased (step S412). If M + L <A with respect to the total length A of the input kana character string, it is determined that the reading length L can be increased, and the process of increasing the reading by one character (ie, incrementing L by 1) Process) (step S414). After increasing the reading length L by the value 1, the above-described processing is repeated from step S405.
[0075]
As a result, all words of reading from the length 1 to the maximum length at the start position M are expanded in the expansion buffer. Word expansion and cost assignment are performed in the same manner as in the first embodiment (see FIG. 4). When the length of the reading is sequentially increased and the word length cannot be increased (step S412), it is then determined whether or not the word search start position M can be moved toward the end of the input kana character string. (Step S416). If it is possible to move, it is determined that all the search for the word starting from the start position M is completed, and after the process of incrementing the start position M by 1 (step S418), the reading length L is set to the value. Return to 1 and repeat or pass the above process. Therefore, when these processes are performed, all the representative words and attached words that may constitute the input kana character string are expanded in the expansion buffer, and the representative words with the derived notation are expanded. A mark is added to this.
[0076]
With the above process, the word search process (step S400 in FIG. 14) is completed. Therefore, next, a segment segmentation process is performed (step S420). Phrase segmentation is well-known, and the combination of phrases is determined using the above-described words expanded in the expansion buffer so that the sum of the values assigned to each word becomes the smallest value. There is no particular difference from the first embodiment regarding the phrase segmentation process.
[0077]
Next, a dependency verification process is performed (step S430). The contents of the dependency test are the same as those in the first embodiment, but in this embodiment, as described above, only representative words are expanded in the expansion buffer, and derived words are not expanded. In the first embodiment, for example, when there is a dependency of “rule” + “ga” + “change”, if there is a derived expression “change” for “change”, “the rule changes” in the dependency dictionary. And the dependency that "the rules change" were memorized. In the present embodiment, since the dependency test itself is performed only by the representative notation, the dependency test for the character string “Kisoku” + “ga” + “alternate” is “rule” + “ga” + “ It is done only for “change”. Therefore, the time required for checking the dependency is shortened. In this embodiment, the dependency dictionary is included in the independent word dictionary 58 and stored in the management area of the semantic example. An example of the dependency dictionary is shown in FIG. The contents of the dependency dictionary include readings, received words, dependency words, and allowed attached words as in the first embodiment (see FIG. 7). Further, the same rules as those in the first embodiment are applied to the determination of the dependency on the duty and the passive and the rule of the dependency (see FIG. 13).
[0078]
In this way, if a dependency test is performed and a word candidate (for example, “change” and “alternate” in “Kisoku-ga-wa”) is found that cannot be determined by phrase writing, a word for which dependency is established is found. Then, the process of setting this word as the first candidate is performed. Thereafter, a process of replacing the word notation is performed for the word that is the first candidate (step S440). In the word replacement process, it is determined whether or not there is a derivation notation for the word that is the first candidate (for example, “change”). The notation used in is examined and replaced with that notation. If the word used immediately before is “changed”, the derived notation “change” is used instead of the representative word that has been used consistently in word search, segmentation processing, and dependency test.
[0079]
FIG. 22 shows a processing routine for performing word replacement processing. In this processing routine, first, the target word to be replaced is set to the word of the first phrase segmented by the segment segmentation process (step S500), and it is determined whether or not the target word is marked (step S510). ). If the mark is attached, it can be determined that the derived notation is present and the replacement to the derived notation is instructed, and the process of replacing the subsequent first candidate with the display word is performed (step S520). The display word is the representative word itself if the representative notation has been used before, and the derived word if the derived notation has been used and learned before. If the target word is not marked, the display word is not replaced.
[0080]
Thereafter, it is determined whether or not an unprocessed word still remains (step S530). If it remains, the target word is shifted backward by one (step S540), and the process is repeated from step S510 described above. If no target word remains, this processing routine ends. In the above-described processing, when there is a derived notation, learning which notation is used as a display word is easy if the word existing at the head of the management area of the derived word is used as the display word. . In this case, when using a representative word, the representative notation itself (or derivative notation information corresponding to the representative notation) may be stored at the beginning of the derived word, or the representative notation may be stored at the beginning of the derived word. A flag indicating whether to use or use a derived notation may be stored. Further, when using a derivation notation, it is also possible to store the notation in the representative word management area.
[0081]
In the display process following the word replacement process, a process of displaying the word candidates after the kana-kanji conversion on the CRT 26 is performed in accordance with the notation of the replaced word. As shown in FIG. 19 and FIG. 20, the derivation notation does not prepare kanji corresponding to the derivation notation, but records a number indicating the type of the derivation notation after the derivation notation separator. Only. Therefore, in the display process, according to this number, for example, if the derivative notation information is “4”, it is determined that the feed is “permitted” and the representative word “change” is displayed as “change”.
[0082]
According to the present embodiment described above, only candidate words are used as word candidates until a word is searched for the input kana character string, the dividing process is performed, and the dependency process is performed. Therefore, even when a word for which derivation notation is permitted is searched, it is not necessary to consider the derivation notation, and each process can be performed at high speed. In addition, the capacity of the expansion buffer for expanding word candidates can be reduced. Furthermore, even when there are a plurality of dependencies, there is an advantage that the determination of the dependencies can be easily performed. This will be described with reference to the example of FIG. As illustrated in FIG. 21, there are three types of dependency of “rule” + “ga” + “change”, “handling” + “ga” + “change”, “luggage” + “no” + “handling”. It is assumed that there is a derivative notation of “change” and “handling” for each word. In this case, for example, it is assumed that the expression “rule” + “ga” + “change” is learned for “the rule changes”. In this case, if “change” and “change” are managed separately as dependency information, “change” is learned in the conversion of “replacement” and “conversion”. However, it is converted to “handling” + “ga” + “change”. On the other hand, in the present embodiment, learning that “change” is to use “allowance” in the derived notation instead of “main rule” is made, so “change” even if another dependency is established. For, "change" will be used consistently.
[0083]
In addition, when converting “the handling of a child”, in the conventional dependency determination, even if there is dependency information “package / handling”, there is a dependency “package / handling”. If it does not exist, it will compare the establishment of the chain of dependency of "package / handling / changing" with the formation of the chain of "handling / changing" and give priority to the formation of the former dependency Become. As a result, even if “handling / change” is learned immediately before, it is converted to “package / handling / change”. On the other hand, in the present embodiment, since it is only learned to use the derivation of “acceptable” for the word “handling”, once “handling” is learned, the “handling / handling of baggage / changes” correctly. Will be converted.
[0084]
In this embodiment, the word search, the phrase segmentation process, and the dependency test are performed using the representative word. However, a part of these processes may be performed using the representative notation and the derived notation. Absent.
[0085]
Although several embodiments of the present invention have been described above, the present invention is not limited to these embodiments. For example, instead of the minimum cost method, another phrase segmentation method such as a two-phrase longest match method may be used. It goes without saying that the present invention can be implemented in various modes within a range that does not depart from the gist of the present invention, such as the configuration used.
[0086]
【The invention's effect】
As described above, in the first kana-kanji conversion device and the kana-kanji conversion method of the present invention, when a phrase corresponding to the information of the related word and the received word is searched, the word attached to the related word is an allowable attached word. If the attached word following the received word is a word representing a use or passive, it is determined whether the word attached to the related word is a word corresponding to the use or passive. Limit phrase segmentation candidates based on search results. Therefore, when there is dependency information, instead of simply looking at the establishment of the dependency due to the presence of both words, it is assumed that the dependency attached when the word attached to the dependency language allows the relationship between the two words, If the received word is a word indicating a service or passive, and if the word attached to the dependency word is a word corresponding to the service or passive, the phrase candidate is limited as the formation of the dependency. An undesired candidate for split writing becomes difficult to select, and there is an excellent effect that the possibility of desired split writing is increased. If the priority order of kanji candidates for each phrase is to be changed on the premise of phrase segmentation that has already been estimated by another method instead of the restriction of phrase candidates, the desired kanji candidate is set as the first candidate. The possibility of being obtained can be increased. Not only can you change the priority order of word candidates using dependency information, but it can also be used for dependency information at the segmental writing stage, and it can be determined whether the dependency is established in the case of use or passive. It is possible to obtain a handwriting candidate that makes use of dependency, which is a higher level language activity. In addition, it is not necessary to prepare a special dictionary for analysis of usage / passive dependency, and processing can be performed at high speed.
[0087]
In addition, according to the kana-kanji conversion device of claim 2, since only the representative notation is used in the dependency test, the time required for the dependency test and the required memory capacity can be reduced. Further, in the kana-kanji conversion device according to claim 3, since the dependency information including the above-mentioned usage / passive is used to change the priority order of the kanji candidates for each phrase instead of the restriction of the phrase segmentation, the dependency information It is possible to obtain a desired kanji candidate with high priority using. In addition, according to the kana-kanji conversion device of claim 4, there is an effect that the search time for the split writing is short, and the linked dependency is not selected erroneously.
[0088]
According to the kana-kanji conversion device according to claim 5, there is an effect that a phrase candidate can be preferentially made a phrase candidate. Further, according to the kana-kanji conversion device according to claim 6, there is an effect that the phrase segmentation is preferentially selected based on the dependency and the word for which the dependency is established is set as the first candidate for kana-kanji conversion. .
[0089]
According to the kana-kanji conversion device of claim 7, it is possible to reduce the amount of information related to the attached word. According to the kana-kanji conversion device according to claim 8, there is an effect that the dependency can be found even when there is a modification word between the dependency word forming the dependency and the reception word. . In addition, when this dependency relationship is found, if the range from the starting phrase to the found phrase is excluded from the search range of the next dependency, This makes it possible to avoid erroneous determination of correct dependency and to speed up the search for dependency.
[0090]
According to the kana-kanji conversion device of claim 9, even when a phrase including a word having a dependency relationship is not found, an appropriate word candidate can be selected.
[0091]
According to the second kana-kanji conversion device and the kana-kanji conversion method of the present invention, when performing the dependency test with reference to the dependency information dictionary, the test is performed using only the representative notation registered in the word dictionary. The word constituting the input character string is specified. When displaying the candidate character string of the specified word, not only the representative notation stored in the word dictionary but also the derivative notation is used for display. As a result, it is possible to achieve both speeding up of grammar processing including dependency and diversity of notation.
[Brief description of the drawings]
FIG. 1 is a functional block diagram showing an implementation form of a kana-kanji conversion function in a kana-kanji conversion apparatus according to an embodiment of the present invention.
FIG. 2 is a block diagram illustrating hardware that realizes a kana-kanji conversion device as an embodiment;
FIG. 3 is a flowchart showing a phrase segmentation process executed in the segment segmentation unit 102;
FIG. 4 is an explanatory diagram showing a state of phrase segmentation by a minimum cost method.
FIG. 5 is an explanatory diagram showing another example of phrase segmentation by the minimum cost method.
FIG. 6 is an explanatory diagram showing a dependency test process in the embodiment.
FIG. 7 is an explanatory diagram illustrating an example of a dependency dictionary according to the first embodiment.
FIG. 8 is an explanatory diagram showing an example of phrase segmentation processing performed using dependency information.
FIG. 9 is an explanatory diagram showing a state of processing of another sentence example.
FIG. 10 is an explanatory diagram showing priorities of candidates for kana-kanji conversion in that case as well.
FIG. 11 is an explanatory diagram showing a state of different types of dependency processing;
FIG. 12 is an explanatory diagram showing priorities of candidates for kana-kanji conversion in that case.
FIG. 13 is an explanatory diagram showing a type when there are a plurality of dependencies in one input character string.
FIG. 14 is a flowchart showing a kana-kanji conversion processing routine in the second embodiment.
FIG. 15 is a flowchart showing details of the word search processing;
FIG. 16 is an explanatory diagram showing an internal configuration of the independent word dictionary 58;
FIG. 17 is an explanatory diagram showing how a basic word area is managed.
FIG. 18 is an explanatory diagram showing a configuration of word data.
FIG. 19 is an explanatory diagram showing details of word information together with each separator.
FIG. 20 is an explanatory diagram illustrating an example of derivation notation.
FIG. 21 is an explanatory diagram showing a schematic configuration of a dependency dictionary and an example of representative notation and derivative notation;
FIG. 22 is a flowchart showing a word replacement processing routine.
[Explanation of symbols]
21 ... CPU
22 ... ROM
23 ... RAM
24 ... Keyboard
25 ... Keyboard interface
26 ... CRT
27 ... CRTC
28 ... Printer
29 ... Printer interface
30: Hard disk controller (HDC)
31 ... Bus
32 ... Hard disk
40 ... Character input part
42. Conversion control unit
44 ... converted character string output section
50 ... Character string input part
52 ... Character storage
54 ... Independent word candidate creation section
56 ... Independent word analysis position management department
58 ... Autonomous dictionary
64 ... Attached word candidate creation part
66 ... Attached word analysis position management section
68 ... Attached word dictionary
70. Dependency learning department
70 ... Learning Department
72 ... Independent language learning department
74 ... Auxiliary Language Learning Department
76 ... Affix Learning Department
78 ... Character conversion learning unit
80: Word data creation unit
82 ... Connection verification department
84 ... Connection verification table
86 ... Cost calculator
90 ... Dependency candidate adjustment section
92 ... Passive analysis section
94 ... Participant tolerance analysis part
96 ... Dependency range management department
98 ... Dependency dictionary
100: Word data storage unit
102 ... sentence segment writing section
104 ... Dependency transposition information adjustment unit
106 ... phrase data storage unit
108 ... Conversion character string output section

Claims (9)

A kana-kanji conversion device that refers to a word dictionary and creates a phrase kanji writing candidate by segmenting an input kana character string and generating a candidate character string constituting a kana-kanji mixed sentence using the phrase kanji writing candidate There,
A dependency information dictionary that stores information on a dependency word and a reception word that constitute a dependency between predetermined phrases together with information on an allowable adjunct allowed between the dependency word and the reception word;
When performing the process of segmenting the input character string, referring to the independent word dictionary storing the independent words and the auxiliary word dictionary storing the auxiliary words, the words between the words that can be generated from the input kana character string Generating means for generating the combination comprehensively;
A phrase search means for searching a phrase having a word corresponding to the information of the dependency word and the received word with reference to the dependency information from the generated combination of words;
First determination means for determining whether a word attached to the related word is the permissible attached word when a phrase having a word corresponding to the related word and the received word information is searched;
When a clause having a word corresponding to the related word and the received word information is searched, if the attached word following the received word is a word representing a use or passive, the word attached to the related word is A second determination means for determining whether the word corresponds to a use or passive character;
When either of the first and second determination means makes an affirmative determination, the combination of the searched words is regarded as a priority candidate in the phrase segmentation, and when neither of them makes an affirmative determination Kana-kanji conversion device comprising: phrase candidate restriction means that does not make the combination of the searched words a priority candidate in the phrase segmentation. A kana-kanji conversion device according to claim 1,
The dictionary is a word dictionary storing a word reading and a notation corresponding to the reading, and when there are a plurality of notations as a notation corresponding to the word reading, , With a notation information section that stores the notation defined as a derivative notation,
The phrase search means is means for searching for the dependency using only the representative notation stored in the word dictionary,
Furthermore,
A kana-kanji conversion device comprising candidate character string display means for displaying the converted candidate character string using the representative notation and the derived notation stored in the notation information section of the word dictionary for the phrase segmentation candidate. A kana-kanji conversion device according to claim 1 or 2,
The phrase search means includes:
A backward search means for searching for phrases including words corresponding to the dependency information in order forward, excluding the already registered searched range, starting from the backward phrase;
Searched range registration means for registering from the starting phrase to the found phrase as a searched range of dependency information when a phrase having a word corresponding to the dependency information is found by the search Kana-Kanji conversion device with A kana-kanji conversion device according to claim 1 or 2,
The phrase candidate restriction unit is a unit that preferentially selects a phrase segmentation candidate including the word when the first or second determination unit determines the presence of the corresponding quasiword, the received word, and its associated word. Kana-Kanji conversion device with   The phrase search means includes means for determining that the dependency relation is established when the auxiliary word existing between the words having the dependency relation is a word having a predetermined specific grammatical structure. Item 3. A kana-kanji conversion device according to item 1 or 2. A kana-kanji conversion device according to claim 1 or 2,
The phrase search means refers to dependency information stored in the dependency dictionary with a predetermined phrase as a starting point, and a phrase having a word corresponding to the dependency information as a starting phrase An alternative phrase search means for searching to phrases other than
Search range exclusion means for excluding the range from the starting phrase to the found phrase from the search range by the next phrase search means when a dependency relationship is found by the separate phrase search means; Kana-Kanji conversion device with A kana-kanji conversion device according to claim 1 or 2,
With respect to a range in which a phrase including a word having a dependency relationship is not found with reference to the dependency dictionary, means for selecting a combination that maximizes the combination of words and the likelihood of the combination of phrases. Kana-kanji conversion device provided. A kana-kanji conversion method for referring to a word dictionary, writing an input kana character string into phrases, and generating a kana-kanji mixed character string candidate by a computer,
When performing the process of segmenting the kana character string input from the keyboard, it can be generated from the input kana character string by referring to the independent word dictionary storing the independent words and the auxiliary word dictionary storing the auxiliary words A comprehensive combination of words is generated by a computer,
Dependency information is stored by referring to the dependency information stored in the combination of the generated words together with the information on the allowable adjunct allowed between the dependency words. Use a computer to search for phrases with words that correspond to the received information,
When a phrase having a word corresponding to the related word and received word information is searched and found,
Determining whether the word attached to the dependency word is the allowable adjunct;
If the adjunct word following the received word is a word representing a use or passive character, it is determined whether the word attached to the related word is a word corresponding to the use character or passive;
If any one of the determination results is an affirmative determination, the combination of the searched words is set as a preferential candidate in the phrase segmentation, and if any determination result is not an affirmative determination, A kana-kanji conversion method that restricts candidates for phrase segmentation by a computer by not using a combination of searched words as a priority candidate in the segment segmentation. A kana-kanji conversion method according to claim 8,
When there are a plurality of notations as notation corresponding to the word reading in the word dictionary storing the reading of the word and the notation corresponding to the reading, the notation determined as the representative notation and the derived notation are determined. Remember the notation,
For each of the segmented phrases, only the representative notation stored in the word dictionary is used to perform a dependency test using the dependency information dictionary, and the words constituting the input character string are Identified by computer,
A kana-kanji conversion method for displaying a converted candidate character string on a display device using the representative notation and derivative notation stored in the word dictionary for the identified word.

Images